Edinburgh Research Explorer

IL-10-producing, ST2-expressing Foxp3(+) T cells in multiplesclerosis brain lesions

Citation for published version:Zandee, SEJ, O'Connor, RA, Mair, I, Leech, MD, Williams, A & Anderton, SM 2017, 'IL-10-producing, ST2-expressing Foxp3(+) T cells in multiple sclerosis brain lesions', Immunology and Cell Biology.https://doi.org/10.1038/icb.2017.3

Digital Object Identifier (DOI):10.1038/icb.2017.3

Link:Link to publication record in Edinburgh Research Explorer

Document Version:Peer reviewed version

Published In:Immunology and Cell Biology

General rightsCopyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s)and / or other copyright owners and it is a condition of accessing these publications that users recognise andabide by the legal requirements associated with these rights.

Take down policyThe University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorercontent complies with UK legislation. If you believe that the public display of this file breaches copyright pleasecontact openaccess@ed.ac.uk providing details, and we will remove access to the work immediately andinvestigate your claim.

Download date: 22. May. 2021

1

IL-10-producing, ST2-expressing Foxp3+ T cells in multiple sclerosis brain lesions 1

Running title: CD4+Foxp3+ cells in Multiple Sclerosis lesions. 2

Stephanie Elizabeth Johanna Zandee1, Richard Anthony O’Connor1, Iris Mair1, Melanie Dawn 3

Leech, Anna Williams2 and Stephen Mark Anderton1 4

Affiliations: 5

1 MRC Centre for Inflammation Research, The Edinburgh Centre for Multiple Sclerosis Research and 6

Centre for Immunity Infection and Evolution, University of Edinburgh, UK. 7

2 MRC Centre for Regenerative Medicine and The Edinburgh Centre for Multiple Sclerosis Research, 8

University of Edinburgh, UK. 9

10

Correspondence: 11

Stephen M Anderton 12

University of Edinburgh, Centre for Inflammation Research, Queen’s Medical Research Institute 13

47 Little France Crescent, Edinburgh EH16 4TJ United Kingdom 14

steve.anderton@ed.ac.uk 15

Phone: +44 131 242 6589 16

FAX: +44-131-242 6682 17

18

Conflict of interest statement: The authors have declared that no conflict of interest exists. 19

20

This work was supported by grants from the UK Medical Research Council, The Dutch MS Research 21

Foundation and the Scottish Chief Scientist Office. 22

23

2

Abstract 24

CD4+Foxp3+ T regulatory (Treg) cells provide a key defence against inflammatory disease, but also have 25

an ability to produce pro-inflammatory cytokines. The evidence for these two possibilities in multiple 26

sclerosis (MS) is controversial. However, this has largely been based on studies of circulating Treg cells 27

derived from peripheral blood, rather than the central nervous system. We show that Foxp3+ cells in the 28

brains of MS patients predominantly produce IL-10 and show high expression of the IL-33 receptor ST2 29

(associated with potent Treg function), indicating that Treg in the inflamed brain maintain their 30

suppressive function. 31

32

3

Main Text 33

CD4+Foxp3+ Treg cells control immune responses in inflamed tissues as well as secondary lymphoid 34

organs1, 2. Treg cells isolated from the peripheral blood of MS patients are reported to show reduced 35

suppressive function, but not reduced frequencies3-5. Treg cells can “trans-differentiate” to a pro-36

inflammatory function, producing IFN-γ or IL-17, when placed in conducive experimental conditions6-8. 37

Peripheral blood Treg cells from MS patients were reported to display this ability, producing IFN-γ in 38

vitro under the influence of IL-129. The major drawback of such studies is that, out of necessity, only the 39

peripheral blood of MS patients can be sampled and not the central nervous system (CNS) itself. Tissue 40

inflammation can stabilize, rather than diminish Treg suppressive function. We reported that the 41

accumulation of highly activated and suppressive, IL-10-producing Treg cells within the CNS is 42

necessary for the natural resolution of experimental autoimmune encephalomyelitis (EAE), a mouse 43

model of MS10, 11. In addition, these CNS Treg cells resisted conversion to pro-inflammatory function in 44

vitro12. Here, we sought to understand the distribution of Treg in human MS lesions and to gather 45

evidence for suppressive, or pro-inflammatory roles for these cells. 46

47

Results and Discussion 48

Immunohistochemistry identified the presence of CD4+Foxp3+ T cells in post mortem brain tissue of 7/11 49

secondary progressive MS (SPMS) patients (Fig. 1a), with none found in control brain tissue. MS 50

samples that did or did not contain CD4+Foxp3+ cells could not be distinguished based on patient gender, 51

age, duration of disease, or time to post mortem processing (summarised in Supp Table 1). CD4+Foxp3+ 52

cells were distributed at similar frequencies across different white matter lesion types (9/10 active lesions, 53

3/7 chronic active borders, 3/7 chronic active centres, 9/17 chronic inactive lesions), but not in 54

remyelinating lesions (Fig. 1b). Thus, Treg presence in MS lesions appears to be associated with the 55

presence of an inflammatory infiltrate (not found in remyelinating lesions). This is consistent with our 56

previous EAE data showing that Treg numbers in the CNS decline markedly, in-line with the 57

inflammatory infiltrate, as the disease resolves10,11. Where present, the frequencies of CD4+ cells that 58

4

were Foxp3+ ranged between 10-30% (Fig. 1c), which represents an enrichment over the expected 59

frequencies of these cells amongst CD4+ T cells in human peripheral blood (1-3% in healthy controls and 60

MS patients) and in cerebrospinal fluid (3-4% in MS)3, 4, 13. 61

Although the presence of CD4+IL-17+ T cells has been reported before14, no analogous analysis 62

has been made of cytokine production by Foxp3+ cells in MS lesions. Two-colour immunohistochemistry 63

identified co-expression of Foxp3 with IL-10, IL-17, IFN-γ, or GM-CSF in active and chronic lesions 64

(Fig. 2a). Approximately 50% of Foxp3+ cells stained positive for IL-10 (Fig. 2b). Lower frequencies of 65

Foxp3+ cells stained positive for pro-inflammatory cytokines. IL-10 was the dominant cytokine produced 66

by Foxp3+ cells (>60%) in active lesions and the borders of chronic active lesions (Fig. 2b). This was less 67

evident in chronic inactive lesions and in the centres of chronic active lesions, where Foxp3+ cells showed 68

no enrichment in IL-10 over other cytokines. In contrast to IL-10, TNF-α staining in Foxp3+ cells only 69

became evident in chronic inactive lesions. Frequencies of Foxp3+ cells staining for IFN-γ, IL-17, or GM-70

CSF were low in all active and chronic lesion types. We conclude that the predominant cytokine produced 71

by Foxp3+ cells within the brains of SPMS patients is IL-10. This is entirely consistent with our previous 72

observations of Treg in the CNS of mice with EAE10-12 and indicates that, in MS, Treg that infiltrate the 73

lesions are in suppressive rather than pro-inflammatory mode. 74

As CD4+Foxp3+ cells composed only a minor fraction of infiltrating cells within lesions, their 75

contribution to the overall cytokine+ cells remained modest, even for IL-10. We compared the frequencies 76

of CD4+Foxp3- or CD4+Foxp3+ cells in all lesions, with the overall levels of cytokine+ cells in those 77

lesions. CD4+Foxp3- frequencies did not correlate with any cytokine (Fig. 3a). Nor did CD4+Foxp3+ cells 78

correlate with IFN-γ, GM-CSF or IL-17. However, CD4+Foxp3+ frequencies correlated with the 79

frequencies of total IL-10+ cells and total TNF-α+ cells (Fig. 3b). 80

Elegant murine studies have shown that IL-10 signalling in Treg cells is required for their own 81

IL-10 expression and subsequent suppressive function15. Therefore it is plausible that, in addition to 82

contributing to the IL-10 pool, IL-10+ Treg cells are specifically attracted to, expanded in, or maintained 83

5

in lesions with high IL-10 levels. TNF-α-blockade is a potent therapeutic option for several human 84

inflammatory diseases such as rheumatoid arthritis, Crohn’s disease and psoriasis16, 17, but not MS18. 85

Studies on how TNF-α-blockade effects the Treg populations have led to conflicting results. TNF-α 86

blockers have been reported to increase the number or function of Treg cells in RA and Crohn’s19, 20. 87

However, it has also been shown to inhibit suppressive function of Treg cells through down-regulation of 88

Foxp3 in RA patients21. Recent studies indicate TNF-α signals selectively through TNFR2 in Treg cells22, 89

23. This suggests that Treg cells might require TNF-α for their suppressive function and provides a 90

plausible explanation for the positive correlation between Foxp3+ cells and TNF-α+ cells that we see. 91

Expression of the IL-33 receptor, ST2, has been associated with potent Treg function in murine 92

models24-26. Indeed, we found ST2 to be particularly enriched in CNS Treg in EAE (Fig. 4a). IL-33 is 93

highly expressed in the CNS in both EAE and MS (Fig. 4b)27, 28. Dual immunofluorescence identified the 94

presence of Foxp3+ST2+ cells in MS brains (Fig. 4c). In particular, ~60% of Foxp3+ cells in active lesions 95

were ST2+, whilst its expression was almost absent in Foxp3+ cells in chronic lesions (Fig. 4d). High 96

expression of both IL-10 (Fig. 2b) and ST2 (Fig. 4d) by Foxp3+ cells in active lesions suggests that their 97

suppressive potency should be greatest in these lesions and that this might wane in more chronic lesions. 98

A recent study from Miron et al29 demonstrated high numbers of M2 macrophages, also particularly in 99

active lesions, of the same brain tissue studied here. This is interesting for two reasons. Firstly, IL-10 100

(perhaps originating from Treg cells) can promote the M2 phenotype, which is thought to contribute to 101

remyelination by inducing oligodendrocyte differentiation. Secondly, a study of experimental cerebral 102

malaria recently reported that IL-33 is protective by coordinating both Treg and M2 activity (the latter via 103

expansion of type-2 innate lymphoid cells which release M2-promoting cytokines)30. Whether such a 104

coordinated response is protective in CNS autoimmune inflammation, and whether there are viable 105

therapeutic approaches that can boost the numbers and/or sustain the function of these cells, should be 106

fruitful avenues for exploration. 107

108

6

109

7

Methods 110

Human Tissue specimens 111

Post-mortem tissue from SPMS patients and control individuals who died of non-neurological causes 112

were obtained via a UK prospective donor scheme with full ethical approval and informed consent from 113

the UK Multiple Sclerosis Tissue Bank (MREC/02/2/39)(Supplementary information Table 1). Snap 114

frozen unfixed tissue blocks from 11 SPMS patients (a total of 16 blocks containing 10 active lesions, 7 115

chronic active lesions, 17 chronic inactive lesions and 12 remyelinating lesions) and 4 control blocks were 116

analysed. Lesions were classified as active, chronic active, chronic inactive and remyelinating according 117

to the International Classification of Neurological Diseases (www.icdns.org) using Luxol Fast Blue – 118

Cresyl Violet staining and Oil Red O staining. 119

Immunohistochemistry of T cell subsets 120

10 μM sections were fixed in 4% PFA (Fisher Scientific, Waltham, USA) and subsequently delipidised in 121

70% ice-cold ethanol. Antigens were retrieved using heating in acid citric buffer (Vector, Burlingame, 122

USA). Sections were incubated with anti-Foxp3 (ab10563, rabbit, Abcam, Cambridge, UK) overnight at 123

4°C. Subsequently the sections were incubated with anti-CD4 (M7310, mouse, Dako, Glostrup, Denmark) 124

for 30 minutes at room temperature. An EnVision G|2 Doublestain System, Rabbit/Mouse kit (Dako) was 125

used for detection as per manufacturer’s instructions, with exception of the use of an Vector Blue 126

Alkaline Phosphatase Substrate Kit III (Vector) to develop the signal. Sections were mounted in aqueous 127

permafluor medium (Thermo Scientific, Waltham, USA). Primary antibodies were omitted to check for 128

non-specific binding of polymers. Rabbit IgG (ab27478, Abcam) or Mouse IgG1 isotype control (X0931, 129

Dako) were used to control for non-specific binding of the primary antibodies. All IHC experiments were 130

performed in triplicate. 131

132

133

8

Immunohistochemistry of cytokines 134

For double staining of Foxp3 and cytokines, combinations of antibodies against TNF-α, IFN-γ, IL-17, 135

GM-CSF or IL-10 (AF-210-NA, AF-285-NA, AF-317-NA, AF-215-NA, AF-217-NA, all goat, R&D 136

systems, Abingdon, UK) with anti-Foxp3 (rabbit, Abcam) were used. For single IL-33 staining a goat 137

anti-IL-33 antibody (AF3625, R&D systems) was used. Briefly, frozen brain sections were fixed in 4% 138

PFA (Fisher Scientific), followed by antigen retrieval as described above. Endogenous peroxidase was 139

blocked with 3% H2O2 in dH2O (Fisher Scientific), followed by blocking of biotin for 15 minutes 140

(Vector). Sections were incubated with 10% horse serum in PBS (Biosera, Boussens, France) and Fc 141

Receptor Blocking Solution was added (Human TruStain FcX Biolegend, London, UK). Primary 142

antibodies were added overnight at 4°C. Cytokines were detected with donkey anti-goat-biotin (ab6578, 143

Abcam) followed by streptavidin-alkaline phosphatase (SA-5100, Vector) and visualized with the Vector 144

Blue Alkaline Phosphatase Substrate Kit III (Vector). Slides were blocked with 10% goat serum in PBS 145

(Biosera). Anti-Foxp3 (rabbit) was detected with an anti-rabbit polymer-HRP (Dako) and developed with 146

DAB substrate (Dako). Sections were counterstained with 4',6-diamidino-2-phenylindole (DAPI) (Life 147

Technologies, Carlsbad, USA), and mounted in aqueous permafluor medium (Thermo Scientific). 148

Secondary antibodies/polymers alone, or normal goat IgG (AB-108-C, R&D Systems) and rabbit IgG 149

(Abcam) were used to control for non-specific binding. 150

151

Immunofluorescent staining of Foxp3 and ST2 152

Sections were air dried overnight, fixed in ice-cold acetone (VWR) and air dried for 30 minutes. 153

Endogenous peroxidase and biotin were blocked as described above. Sections were blocked with 10% 154

goat serum (Biosera) in PBS and incubated with rabbit anti-Foxp3 (Abcam) overnight at 4oC. Foxp3 155

antibody was detected with a goat-anti-rabbit-biotinylated antibody (BA-1000, Vector), followed by 156

incubation with a streptavidin-coupled horseradish peroxidase (SA-5004, Vector). Tyramide-Cy3 (Perkin-157

Elmer, Waltham, USA) was applied for 10 minutes to visualize the staining and ST2L FITC antibody 158

(MdBioproducts, Zürich, Switzerland) was incubated overnight at 4oC. Sections were counterstained with 159

9

DAPI (Life Technologies) and mounted in aqueous Permafluor medium (Thermo Scientific). Mouse 160

IgG1 FITC (1053002F, MdBioproducts), rabbit IgG (Abcam), or secondary antibodies/polymers alone 161

were used to control for non-specific binding. Only lesions with Foxp3+ cells were analysed. 162

163

EAE induction 164

C57BL/6 mice were bred under specific pathogen free conditions at the University of Edinburgh. All 165

experiments were approved by the University of Edinburgh Ethical Review Committee and were 166

performed in accordance with UK legislation. Female mice were used between 6-12 weeks old (n = 7). 167

EAE was induced by administration 100μg of MOG35–55 peptide (MEVGWYRSPFSRVVHLYRNGK, 168

Cambridge Research Biochemicals, Teesside, UK), emulsified in complete Freund’s adjuvant containing 169

200μg of heat-inactivated Mycobacterium tuberculosis H37Ra (Sigma-Aldrich), with a total volume of 170

100μl injected subcutaneously into the hind legs. On the same day and 48 hours later, 200ng of pertussis 171

toxin (Health Protection Agency, Dorset, UK) was given in 0.5ml of PBS intraperitoneally. Clinical signs 172

of EAE were assessed daily with the following scoring system: 0, no signs; 1, flaccid tail; 2, impaired 173

righting reflex and/or gait; 3, partial hindlimb paralysis; 4, total hindlimb paralysis; 5, hindlimb paralysis 174

with partial forelimb paralysis; 6, moribund or dead. 175

176

Isolation of CNS mononuclear cells and flow cytometry 177

Mice were sacrificed at d16 (when Treg were evident in the CNS) by CO2 asphyxiation and perfused with 178

PBS. Brains and spinal cords were removed, mechanically disrupted and digested in RPMI containing 7.5 179

mg/ml collagenase type 4 (Lorne Laboratories, Reading, UK) and 2.5 mg/ml DNAse I (Sigma-Aldrich) 180

for 30 minutes at 37°C. Mononuclear cells were isolated from the interface of a 30%:70% discontinuous 181

Percoll gradient (GE healthcare, Uppsala, Sweden) after centrifugation at 530xg for 20 minutes. Cells 182

were stained using the following antibodies: anti-CD4 brilliant violet 650 (Biolegend), anti-Foxp3 eFluor 183

450 (eBioscience, San Diego, USA), anti-ST2 FITC (MdBioscience). 184

185

10

Data acquisition 186

Immunohistochemistry samples were analysed using an Olympus AX70 microscope (Olympus 187

Corporation, Tokyo, Japan). The number of cells was always quantified in the whole lesion and expressed 188

as cells per mm2 within different lesion types. The total number of nuclei was also documented. An 189

AxioScan.Z1 slide scanner (Zeiss, Cambridge, UK) was used to acquire fluorescent images and Zen Blue 190

software (Zeiss) used to process the fluorescent images. Experiments were repeated 2-3 times and 191

analysed blinded. Flow cytometric data was acquired using a Becton Dickinson (BD, Franklin Lakes, 192

USA) LSRFortessa II and analysed using FlowJo software (Tree Star version 3.2.1, Ashland, USA). 193

194

Statistical analysis 195

Where data were unevenly distributed, log transformations and statistical analysis was performed using a 196

linear mixed model. This model accounts for random effects such as having different numbers of tissue 197

blocks from each patient. In case of multiple testing, significant values were corrected with the 198

Bonferroni test. When random effects were found to be non-significant, simplified statistical tests such as 199

a Mann-Whitney-U test or a Kruskal-Wallis test were used. In case of multiple testing using a Kruskal-200

Wallis test, significant values were corrected with Dunn’s multiple comparison test. Correlations were 201

performed using Spearman rank correlation tests. Lesions were not subdivided into pathological types, 202

thereby allowing sufficient numbers for analysis. SPSS version 19 (IBM, New York, USA) statistical 203

software and Prism version 5.04 (Graphpad, La Jolla, USA) software were used to perform the 204

calculations. Data are presented as mean ± SEM. Significant differences are denoted as * p<0.05, ** 205

p<0.01 and *** p<0.001. 206

207

Acknowledgements: We thank the UK Multiple Sclerosis Tissue Bank for providing human brain tissue, 208

F. Roncaroli (Imperial College London) for neuropathological diagnosis, R. Nicholas (Imperial College 209

London) for providing clinical histories, Mr. D.M. Mole for collaboration in developing several 210

immunohistochemistry techniques, Ms. A. Boyd for technical assistance, Dr. M. Chase-Topping for 211

11

statistical advice and staff of Flow Cytometry and Histology/Imaging Facilities. This work was supported 212

by the Scottish Chief Scientist Office (ETM/163), the Dutch MS Foundation (r12-1MS) and the UK 213

Medical Research Council (G0801924). 214

215

Conflicts of interest: The authors have no conflicting interests to declare. 216

217

12

Supplementary information is available at the Immunology and Cell Biology website. 218

219

13

References 220

1. Sakaguchi S, Fukuma K, Kuribayashi K, Masuda T. Organ-specific autoimmune diseases induced in mice 221 by elimination of T cell subset. I. Evidence for the active participation of T cells in natural self-tolerance; deficit of a 222 T cell subset as a possible cause of autoimmune disease. J Exp Med 1985; 161: 72-87. 223 2. Bennett CL, Christie J, Ramsdell F, Brunkow ME, Ferguson PJ, Whitesell L, et al. The immune 224 dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX) is caused by mutations of FOXP3. Nat 225 Genet 2001; 27: 20-21. 226 3. Viglietta V, Baecher-Allan C, Weiner HL, Hafler DA. Loss of functional suppression by CD4+CD25+ 227 regulatory T cells in patients with multiple sclerosis. J Exp Med. 2004; 199: 971-979. 228 4. Haas J, Hug A, Viehover A, Fritzsching B, Falk CS, Filser A, et al. Reduced suppressive effect of 229 CD4+CD25high regulatory T cells on the T cell immune response against myelin oligodendrocyte glycoprotein in 230 patients with multiple sclerosis. Eur J Immunol 2005; 35: 3343-3352. 231 5. Anderton SM, Liblau RS. Regulatory T cells in the control of inflammatory demyelinating diseases of the 232 central nervous system. Curr Opin Neurol 2008; 21: 248-254. 233 6. Xu L, Kitani A, Fuss I, Strober W. Cutting edge: regulatory T cells induce CD4+CD25-Foxp3- T cells or 234 are self-induced to become Th17 cells in the absence of exogenous TGF-beta. J Immunol 2007; 178: 6725-6729. 235 7. Zhou X, Bailey-Bucktrout SL, Jeker LT, Penaranda C, Martinez-Llordella M, Ashby M, et al. Instability of 236 the transcription factor Foxp3 leads to the generation of pathogenic memory T cells in vivo. Nat Immunol 2009; 10: 237 1000-1007. 238 8. Voo KS, Wang YH, Santori FR, Boggiano C, Wang YH, Arima K, et al. Identification of IL-17-producing 239 FOXP3+ regulatory T cells in humans. Proc Natl Acad Sci USA 2009; 106: 4793-4798. 240 9. Dominguez-Villar M, Baecher-Allan CM, Hafler DA. Identification of T helper type 1-like, Foxp3+ 241 regulatory T cells in human autoimmune disease. Nat Med 2011; 17: 673-675. 242 10. McGeachy MJ, Stephens LA, Anderton SM. Natural recovery and protection from autoimmune 243 encephalomyelitis: contribution of CD4+CD25+ regulatory cells within the central nervous system. J Immunol 244 2005; 175: 3025-3032. 245 11. O'Connor RA, Malpass KH, Anderton SM. The inflamed central nervous system drives the activation and 246 rapid proliferation of Foxp3+ regulatory T cells. J Immunol 2007; 179: 958-966. 247 12. O'Connor RA, Floess S, Huehn J, Jones SA, Anderton SM. Foxp3(+) Treg cells in the inflamed CNS are 248 insensitive to IL-6-driven IL-17 production. Eur J Immunol. 2012; 42: 1174-1179. 249 13. Feger U, Luther C, Poeschel S, Melms A, Tolosa E, Wiendl H. Increased frequency of CD4+ CD25+ 250 regulatory T cells in the cerebrospinal fluid but not in the blood of multiple sclerosis patients. Clin Exp Immunol 251 2007; 147: 412-418. 252 14. Tzartos JS, Friese MA, Craner MJ, Palace J, Newcombe J, Esiri MM, et al. Interleukin-17 production in 253 central nervous system-infiltrating T cells and glial cells is associated with active disease in multiple sclerosis. Am J 254 Pathol 2008; 172: 146-155. 255 15. Chaudhry A, Samstein RM, Treuting P, Liang Y, Pils MC, Heinrich JM, et al. Interleukin-10 signaling in 256 regulatory T cells is required for suppression of Th17 cell-mediated inflammation. Immunity 2011; 34: 566-578. 257 16. Feldmann M, Maini RN. Anti-TNF alpha therapy of rheumatoid arthritis: what have we learned? Annu Rev 258 Immunol 2001; 19: 163-196. 259 17. MacDonald TT, Hutchings P, Choy MY, Murch S, Cooke A. Tumour necrosis factor-alpha and interferon-260 gamma production measured at the single cell level in normal and inflamed human intestine. Clin Exp Immunol 261 1990; 81: 301-305. 262 18. van Oosten BW, Barkhof F, Truyen L, Boringa JB, Bertelsmann FW, von Blomberg BME, et al. Increased 263 MRI activity and immune activation in two multiple sclerosis patients treated with the monoclonal anti-tumor 264 necrosis factor antibody cA2. Neurology 1996; 47: 1531-1534. 265 19. Nadkarni S, Mauri C, Ehrenstein MR. Anti-TNF-alpha therapy induces a distinct regulatory T cell 266 population in patients with rheumatoid arthritis via TGF-beta. J Exp Med 2007; 204: 33-39. 267 20. Ricciardelli I, Lindley KJ, Londei M, Quaratino S. Anti tumour necrosis-alpha therapy increases the 268 number of FOXP3 regulatory T cells in children affected by Crohn's disease. Immunology 2008; 125: 178-183. 269 21. Valencia X, Stephens G, Goldbach-Mansky R, Wilson M, Shevach EM, Lipsky PE. TNF downmodulates 270 the function of human CD4+CD25hi T-regulatory cells. Blood 2006; 108: 253-261. 271 22. Chen X, Baumel M, Mannel DN, Howard OM, Oppenheim JJ. Interaction of TNF with TNF receptor type 272 2 promotes expansion and function of mouse CD4+CD25+ T regulatory cells. J Immunol 2007; 179: 154-161. 273

14

23. Chen X, Wu X, Zhou Q, Howard OM, Netea MG, Oppenheim JJ. TNFR2 is critical for the stabilization of 274 the CD4+Foxp3+ regulatory T. cell phenotype in the inflammatory environment. J Immunol 2013;190:1076-1084. 275 24. Schiering C, Krausgruber T, Chomka A, Frohlich A, Adelmann K, Wohlfert EA, et al. The alarmin IL-33 276 promotes regulatory T-cell function in the intestine. Nature 2014; 513: 564-568. 277 25. Matta BM, Lott JM, Mathews LR, Liu Q, Rosborough BR, Blazar BR, et al. IL-33 is an unconventional 278 Alarmin that stimulates IL-2 secretion by dendritic cells to selectively expand IL-33R/ST2+ regulatory T cells. J 279 Immunol 2014; 193: 4010-4020. 280 26. Siede J, Frohlich A, Datsi A, Hegazy AN, Varga DV, Holecska V, et al. IL-33 Receptor-Expressing 281 Regulatory T Cells Are Highly Activated, Th2 Biased and Suppress CD4 T Cell Proliferation through IL-10 and 282 TGFbeta Release. PLoS One 2016; 11: e0161507. 283 27. Christophi GP, Gruber RC, Panos M, Christophi RL, Jubelt B, Massa PT. Interleukin-33 upregulation in 284 peripheral leukocytes and CNS of multiple sclerosis patients. Clin Immunol 2012; 142: 308-319. 285 28. Jiang HR, Milovanovic M, Allan D, Niedbala W, Besnard AG, Fukada SY, et al. IL-33 attenuates EAE by 286 suppressing IL-17 and IFN-gamma production and inducing alternatively activated macrophages. Eur J Immunol 287 2012; 42: 1804-1814. 288 29. Miron VE, Boyd A, Zhao JW, Yuen TJ, Ruckh JM, Shadrach JL, et al. M2 microglia and macrophages 289 drive oligodendrocyte differentiation during CNS remyelination. Nat Neurosci 2013; 16: 1211-1218. 290 30. Besnard AG, Guabiraba R, Niedbala W, Palomo J, Reverchon F, Shaw TN, et al. IL-33-mediated 291 protection against experimental cerebral malaria is linked to induction of type 2 innate lymphoid cells, M2 292 macrophages and regulatory T cells. PLoS Pathog 2015; 11: e1004607. 293 294

15

Figure legends 295

Figure 1. CD4+Foxp3+ T cell are enriched within MS lesions. 296

(a) Representative images from different MS cases (A = active, CA = chronic active and CI = chronic 297

inactive lesions) of immunohistochemistry for CD4 (blue) and Foxp3 (brown). No staining was observed 298

using isotype controls or secondary antibodies alone. Scale bars 20 µm. Accompanying images show 299

lesions (LFB = Luxol Fast Blue – Cresyl Violet). Dotted line represents lesion border. Black boxes 300

delineate the areas CD4+Foxp3- and CD4+Foxp3+ cells were pictured. Scale bars 200 µm. (b) Densities of 301

CD4+Foxp3- and CD4+Foxp3+ cells in the indicated SPMS lesion types. (c) Frequencies of CD4+ cells that 302

were Foxp3+ in the indicated SPMS lesion types. Graphs show means ± SEM. Kruskal-Wallis tests with 303

Dunn’s multiple comparison correction were used. * p<0.05, ** p<0.01. 10 active lesions, 7 chronic 304

active lesions, 17 chronic inactive lesions and 12 remyelinating lesions were studied. 305

306

Figure 2. Foxp3+ cells predominantly produce IL-10 in MS lesions. 307

(a) Representative images of immunohistochemistry for individual cytokines (blue) and Foxp3 (brown). 308

No staining was observed using isotype controls or secondary antibodies alone. Scale bars 20 µm. (b) 309

Frequencies of Foxp3+ cells co-staining for individual cytokines in the indicated SPMS lesion types. 310

Graphs show means ± SEM. A Kruskal-Wallis test with Dunn’s multiple comparison correction was used. 311

** p<0.01. 5 active lesions, 3 chronic active lesions and 8 chronic inactive lesions were studied. 312

313

Figure 3. Frequencies of CD4+Foxp3+ cells correlate with IL-10 and TNF-α levels in MS lesions 314

Relationships between the frequencies of CD4+Foxp3- cells (a), or CD4+Foxp3+ cells (b), and the 315

frequencies of all cells staining for the indicated cytokine. Non-parametric 2-sided Spearman correlations 316

were used. Lesions were not segregated based on pathological type. 10 active lesions, 7 chronic active 317

lesions and 17 chronic inactive lesions were studied. 318

319

Figure 4. Foxp3+ST2+ Treg are present in MS lesions. 320

16

(a) Representative flow cytometry plots (gated on CD4+ cells) and summary data showing the expression 321

of ST2 in CD4+Foxp3+ cells in spleen, lymph nodes (LN) and CNS isolated from mice 16 days after 322

induction of EAE. A one-way ANOVA with Bonferroni’s post test was used. Graphs show means ± 323

SEM. 7 mice were studied. (b) Representative immunohistochemistry image of IL-33 (brown) and 324

haematoxylin (blue) and summary data showing percentage of IL-33+ cells in the indicated human SPMS 325

lesions. No staining was observed using isotype controls. Scale bars 20 µm. (c) Representative 326

immunofluorescent staining for DAPI (blue), ST2 (green) and Foxp3 (red) in an active lesion. Arrows 327

delineate ST2+Foxp3+ cells (insets). No staining was observed using isotype controls. Scale bars 40 µm. 328

(d) Frequencies of Foxp3+ cells that stained for ST2 in the indicated SPMS lesions. 5 active lesions, 2 329

chronic active lesions and 6 chronic inactive lesions were studied. A Kruskal-Wallis test with Dunn’s 330

multiple comparison correction was used. * p<0.05. 331

332